Descripció del projecte
Project Summary
Due to the fact that high volumes of textual content need to be translated within very tight deadlines, Machine Translation is a good solution, especially for delimited domains and for specific pairs of languages. If the limitations to a translation project are not only the textual volume and the delivery time, but also technical limitations, such as the format of files, an LSP has an important loss on human and financial resources at the end these kind of projects.
This is the reason why a controlled language at the stage of content creation, along with a customized MT system and finally coupled with a satisfactory post-editing strategy. and QA and TQA methods, should result in an almost perfect chain of production delivering high quality results both for the customer and for the LSP.
Project Details
In an LSP, even if it is an LSP with large experience on handling multilingual content and adapting our client’s digital strategy worldwide, there are more and more large projects that might not be financially interesting for the client nor for the company because of the price: volume of the sources files and price per word using human translation results in tremendous expensive projects for a -very small- web site, for instance.
This project should select a given domain, a given pair of languages anti develop a customized process for the whole chain: content creation with controlled language, best MT choice anti training, customized post-editing guidelines and line-tuned QA method and, eventually, the establishment of a TQA system in order to follow-up results.
Research Questions
– what is the best type of machine translation to a given domain and a given pair of languages.
– what are the best guidelines with controlled language for copywriting in a given domain and language.
– what are the best trainings and guidelines for light-editing according to the errors obtained from the MT and according to client’s expectations in the defined domain.
These questions should be answered in order to:
– improve cost-efficiency and time-efficiency in an LSP when handling big volume translation projects in that language pair and that domain.
– evaluate how these improvements could be applied to other language pairs or other domains, i.e. establish the most important points of action in order to extrapolate this strategy to similar process but with another subject domain or language pair.
Aims of the Project
This project aims at finding out that the control over the whole chain (from content creation to translation delivery) empowers an LSP as it reduces cost in two ways: less time spent on post-editing from freelancers (external cost), less time spent on mistake correction or changes discussion (internal cost). Therefore, the final goal of establishing a successful method is to provide quality translation products to a client with limited resources (financial or in time) while keeping an interesting profit margin.
